Plant Physiol. (1989) 90, 91 -1 00 Received for publication September 16, 1988 0032-0889/89/90/0010/1 0/$01 .00/0 and in revised form November 4, 1988

Soluble Sugars as the Carbohydrate Reserve for CAM in Pineapple Leaves1

Implications for the Role of Pyrophosphate:6-Phosphofructokinase in Glycolysis

Nancy Wieland Carnal* and Clanton C. Black Department of Biology, San Francisco State University, San Francisco, California 94132 (N.W.C.), and Department of Biochemistry, University of Georgia, Athens, Georgia 30602 (C.C.B.)

ABSTRACT product, oxaloacetate, is reduced to form the store of malic Neutral ethanol-soluble sugar pools serve as carbohydrate acid that accumulates in the vacuoles of chloroplast-contain- reserves for Crassulacean acid metabolism (CAM) in pineapple ing cells of CAM plants. During the day, malic acid is remo- (Ananas comosus (L.) Merr.) leaves. Levels of neutral soluble bilized from the vacuole and decarboxylated. The CO2 so sugars and glucans fluctuated reciprocally with concentrations generated is fixed via the reductive pentose phosphate path- of malic acid. Hexose loss from neutral soluble-sugar pools was way. The 3-carbon product of the decarboxylation reaction, sufficient to account for malic acid accumulation with about 95% either PEP or pyruvate, is converted gluconeogenically to of the required hexose accounted for by tumover of fructose and carbohydrate, thereby replenishing the reserve carbohydrate glucose pools. Hexose loss from starch or starch plus lower pools. The coupling ofreserve carbohydrate turnover, in large molecular weight glucan pools was insufficient to account for noctumal accumulation of malic acid. The apparent maximum part, to the production of large intracellular stores of organic catalytic capacity of pyrophosphate:6-phosphofructokinase (PPi- acids rather than to the synthesis of sucrose for export from PFK) at 150C was about 16 times higher than the mean maximum the cell is a unique feature of CAM. CAM plants commit as rate of glycolysis that occurred to support malic acid accumula- much as 5% to 17% of their total dry weight (e.g. pineapple tion in pineapple leaves at night and 12 times higher than the and Bryophyllum calycinum [22], respectively) to the CAM mean maximum rate of hexose tumover from all carbohydrate cycle. Based upon 02 studies ofCAM plants (e.g. Refs. 1, 20), pools. The apparent maximum catalytic capacity of ATP-PFK at we estimate that the relative quantity of carbohydrate com- 150C was about 70% of the activity required to account for the mitted to support nocturnal acidification in these species is mean maximal rate of hexose tumover from all carbohydrate 20 to 170 times the carbohydrate catabolized at night to meet pools if tumover were completely via glycolysis, and marginally sufficient to account for mean maximal rates of acidification. respiratory energy needs. Therefore, at low night temperatures conducive to CAM and under The current concept of CAM, especially the nature of the subsaturating substrate concentrations, PPi-PFK activity, but not carbohydrate pools that support malate biosynthesis during ATP-PFK activity, would be sufficient to support the rate of gly- the dark period, is based almost entirely on data for species colytic carbohydrate processing required for acid accumulation. of Crassulaceae. In most of these species, starch is the source These data for pineapple establish that there are at least two of hexose for malate synthesis. For at least two species of types of CAM plants with respect to the nature of the carbohydrate Crassulaceae and one of Cactaceae, however, lower mol wt reserve utilized to support nighttime CO2 accumulation. The data glucan pools supplement hexose loss from starch and contrib- further indicate that the glycolytic carbohydrate processing that ute significant quantities of hexose for conversion to PEP, supports acidification proceeds in different subcellular compart- ments in plants utilizing different carbohydrate reserves. and hence, to malate (25, 30). The behavior of low mol wt sugar pools in these species precludes their involvement as carbohydrate sources for malate production (22, 25, 29, 32). In contrast, the present report demonstrates that neutral soluble sugar pools serve as carbohydrate reserves for CAM The nocturnal acidification phase of CAM is supported by in pineapple (Ananas comosus (L.) Merr., Bromeliaceae). the glycolytic conversion of large quantities of reserve carbo- Pronounced diurnal changes in the concentration of ethanol- hydrate to PEP.2 PEP is subsequently carboxylated and the soluble sugars in pineapple leaves were reported by Sideris et al. (23), however, since the carbon fixation pathway utilized 'This research was supported by National Science Foundation by pineapple had not been identified, the significance of the grant DMB-84-0633 1. N.W.C. was supported by an Elizabeth Adams- observation was not apparent. In the present report we con- Ann Morgan Endowed Fellowship, American Association of Univer- sity Women Educational Foundation. firm the early observation of Sideris et al., identify the low 2Abbreviations: ATP-PFK, nucleoside triphosphate dependent 6- mol wt sugars that fluctuate diurnally, characterize the dy- phosphofructokinase; PEP, phosphoenolpyruvate; PPi-PFK, pyro- namics of pool size changes of ethanol-soluble sugars, starch, phosphate dependent 6-phosphofructokinase (pyrophosphate:6-phos- and low mol wt glucans, and interpret the results with respect phofructose 1-phosphotransferase); gfw, gram fresh weight. to the operation ofCAM in pineapple. We show that although 91 92 CARNAL AND BLACK Plant Physiol. Vol. 90,1989 glucan fractions in pineapple leaves fluctuate diurnally, fruc- Tissue Sampling and Extraction tose and glucose pools provide the majority, if not all, the hexose required to support nocturnal CO2 accumulation. A Three successively aged leaves were sampled for each of six preliminary report (2) of these results has been presented. plants every 3 h for a 24-h period. Samples were obtained by The turnover of neutral soluble sugar reserves to support taking successive transverse cuts beginning 15 cm from the CO2 assimilation and malic acid accumulation at night has leaf tip and ending approximately 35 cm from the tip. The important implications with respect to the nature and the composite sample of three leaf sections per plant per time compartmentation ofthe glycolytic reactions that mediate the point weighed about 1 g. Control experiments showed that transfer ofcarbon between these two pools. Plant cells contain starch, titratable acid, and soluble sugar levels in samples cytoplasmic and plastid isozymes of nucleoside-triphosphate- taken at the extremes of similar sample areas early and late dependent phosphofructokinase (10, 16) and a cytoplasmi- in the light period were comparable. Likewise, acid and cally localized PPi-dependent phosphofructokinase (e.g. Refs. carbohydrate levels in the three leaves were similar. 2, 4, 7, and 19). These enzymes catalyze the first unique step The leaf pieces were washed, blotted dry, halved longitudi- ofglycolysis, i.e. the phosphorylation of Fru-6-P to form Fru- nally, then separated into two equivalent subsamples. One 1,6-P2, using ATP or PPi, respectively, as the energy source subsample was stored at -20°C and used later for titratable and phosphate donor. acid determination and malate assay. The remaining subsam- Extractable activities of PPi-PFK from pineapple leaves are ple was immediately sliced into small pieces and homogenized 15 to 20 times higher than activities of ATP-PFK. PPi-PFK in about 5 mL 80% ethanol/g tissue. The homogenate was activities in CAM plants of four other families likewise are quantitatively transferred to a centrifuge tube (final higher (4-70-fold) than respective activities of ATP-PFK (5). ethanol:tissue volume was at least 10:1) and quickly brought The properties of ATP-PFK in plants suggest that it is an to 75°C in a water bath. Particulate material was ground important regulatory enzyme and that it contributes substan- against the tube wall throughout the 5 min extraction period. tially to control of carbon flow through glycolysis (27). The The extract was centrifuged for 3 min at 1400g. The super- high activities of PPi-PFK with respect to ATP-PFK in pine- natant was removed and the pellet reextracted three additional apple and other CAM plants, however, suggest that PPi-PFK times with 80% ethanol at 75°C for 5 min with grinding as participates in glycolytic carbon catabolism in these plants before. The supernatants from the hot 80%-ethanol extrac- especially at night when demand for carbon processing via tions plus two 80%-ethanol washes of the pellet were com- glycolysis to support CO2 accumulation, respiratory energy bined and centrifuged at 22,000g at 4°C for 20 min. The needs, the production of required metabolites and the biosyn- composite supernatant volume was adjusted to 25 mL with thesis ofexportable carbohydrates is high. To test this hypoth- 80% ethanol. Control experiments indicated more than 99% esis, we determined the range of mean maximum rates of of the 80% ethanol-soluble sugars were extracted with this glycolysis in pineapple leaves during the acidification phase procedure. Ethanol-insoluble materials from all centrifuga- ofCAM by measuring nocturnal changes in carbohydrate and tions were combined and resuspended in 5 mL of water. malic acid pools, respectively. These estimates of hexose Ethanol-soluble and insoluble fractions were stored at -20°C. processing via glycolysis were then compared to the catalytic capacities of PPi-PFK and ATP-PFK in pineapple leaves. Glucan Extraction and Starch Precipitation Ethanol-insoluble fractions were thawed, then boiled with MATERIALS AND METHODS refluxing for 15 min. After cooling, 52% HC104 was added to Plant Material give a final concentration of29.4% (13). Samples were ground against the tube wall for 5 min, then intermittently for the Isogenic Ananas comosus (L.) Merr. plants were grown in next 15 min. The extract was centrifuged at 12,000g for 30 the Botany Department greenhouse, University of Georgia, min. The pellet was reextracted in 6.75 mL of 29.4% HC104 Athens, GA. Plants were transferred to a Conviron controlled with grinding as before. Supernatants from the two perchlo- environment chamber three to four weeks before use and rate extractions and two distilled water washes of the pellet acclimated to a 15 h/30'C/60 to 70% RH day and 9 h/ 1 50C/ were combined and the volume adjusted to 25 mL with 90% RH night. Quantum flux density at canopy height was distilled water. The HCl04-solubilized carbohydrate was 250 /AE/m2 s. Plants were fertilized once a week with a termed total glucan. commercial 15:15:15 (N:P205:K20) fertilizer at a rate of 1.8 Iodine-precipitable glucans (herein defined as starch) were g/L. precipitated from a 10-mL aliquot of each HC104 extract as described by Hassid and Neufeld (13). Ethanolic NaCl washes Reagents were repeated twice. The starch precipitate was resuspended in hot water and the volume adjusted to 3 mL after cooling. Perchloric acid, starch, and sucrose were purchased from J. In control experiments, mean recovery (±SE) of added au- T. Baker. Starch was dialyzed against distilled water for 24 h thentic starch aliquots with this procedure was 98% ± 2% with three changes of water and dried in a desiccator. D- (n = 6). Glucose and resorcinol were obtained from Fisher Scientific Total glucan and starch were determined as glucose equiv- Co. Anthrone, glucose analysis kit No. 510, f,-D-fructose, alents in the HC104 extract and the starch preparation, re- invertase, and L-malate were obtained from Sigma Chemical spectively, using 0.1% anthrone in 72% H2SO4 (11, 13). Co. Samples were diluted, if necessary, with 20.3% HC104 and CARBOHYDRATE RESERVES DURING CAM 93 distilled water, respectively. Samples and anthrone reagent KMnO4 in 3% H2S04, and with modified Dische's reagent were precooled on ice. After mixing, assay tubes were kept on (3). ice 10 min, then immersed in a 100°C bath for 7.5 min, then rapidly cooled in an ice bath. A620 was determined after Determination of Titratable Acidity samples warmed to room temperature. Glucose standard solutions, prepared in 20.3% HC104 or water, as appropriate, Leaf samples were sliced into small pieces, then ground in were assayed with three dilutions of each sample in the a small volume of C02-free water in a glass mortar. The standard curve range. Data for all glucan fractions were homogenate was quantitatively transferred to a centrifuge expressed as ,umol hexose eq/gfw. tube, brought to a volume of nearly 10 mL, and boiled for 10 min. After cooling, the extract was centrifuged at 12,000g for 10 min. The pellet was washed twice with C02-free water. Sugars Soluble in 80% Ethanol The three supernatants were combined and the volume ad- A 20-mL aliquot of the ethanol-soluble fraction was evap- justed to 25 mL. An aliquot was titrated to pH 8.3 with 4 mm orated under vacuum at 50°C to a small volume. The final NaOH. Data were expressed as lsmol malic acid/gfw by volume of the concentrated sample plus several water washes applying a 2 Aeq/,4mol malic acid conversion. of the flask was about 5 mL. A neutral soluble sugar fraction was prepared from 2 mL ofthe latter extract according to the Malate Determination method of Sutton (25). Glucose, sucrose, and fructose were Malate was determined according to the method of Gut- assayed. Appropriate standards and blanks were assayed with mann and Wahlefeld (12). Duplicate assays were performed three dilutions of each sample in the standard curve range for on each extract. each set of samples. Glucose was determined with a Glucostat kit (Sigma) with Phosphofructokinase the following modifications: 2.5 mL of reagent A" (Sigma) Extraction and Assay were added to a sample appropriately diluted in a final volume ATP-PFK and PPi-PFK were extracted and assayed as of 0.5 mL. The reaction was terminated with 50 ,uL of 4 N described previously (5), but with 5 mm cysteine replacing,- HCI after 20 min at 30°C and the A425 determined. mercaptoethanol. For determinations of Qlo, PPi-PFK was Sucrose was hydrolyzed for 30 min at 30°C in a 0.5 mL assayed in the presence of 0.5 gM Fru-2,6-P2 and the ATP- reaction mixture containing the sample, 25 jmol ofpotassium PFK assay contained 5 mm DTT. The temperature of each acetate/acetic acid buffer pH 4.5, and 20 units of invertase. assay mixture was determined immediately following the Control experiments indicated sucrose hydrolysis was vir- assay. tually complete in 15 min. Total glucose was determined as described above. Glucose derived from sucrose was deter- RESULTS mined by subtraction. Sucrose concentrations were expressed as ,gmol hexose eq/gfw. Diumal Changes in Titratable Acidity and Malic Acid Fructose and fructosans were assayed according to the Concentrations of titratable acids, malic acid and various method of Huber (15). The sensitivity of this assay initially carbohydrates in pineapple leaves were determined every 3 h necessitated fructose determination in concentrated extracts for a 24-h period (Figs. 1 and 2). Malic acid concentrations that had not been purged of ions. Control experiments indi- estimated from titratable acid determinations were slightly cated phosphorylated fructose compounds did not contribute lower, but highly correlated (r2 = 0.96 (n = 18)) with malic significantly to fructose estimates in these samples. After acid levels determined by assay (Fig. 1, A and B). Since the discovering that reducing the resorcinol concentration to 0.1% underestimate of malic acid from titratable acid determina- increased the assay sensitivity 3x, we were able to determine tions was consistent, accurate determination of malic acid fructose concentrations in neutral extracts. The extract frac- changes over any time interval could be determined from tion utilized for fructose analysis for various plants is noted these data. Consequently, only titratable acids were monitored in pertinent figures. in subsequent work. Mean titratable acid levels decreased from approximately 105 ,mol malic acid eq/gfw to 12 ,mol/ Chromatography gfw during the first 8 h of the light period, rose slightly but remained low during the second half of the light period, and The predominant sugars in neutral ethanol-soluble sugar increased to about 100 ,umol/gfw during the dark period (Fig. extracts of pineapple leaves were identified by one-dimen- IA). The rate of malic acid accumulation increased progres- sional paper chromatography. Fifty ,ug of standard authentic sively throughout the dark period. Malic acid accumulation sugars in 80% ethanol were spotted separately and in a mixed late in the dark period was approximately 1.6 times higher preparation on Whatman No. 1 paper along with aliquots of than during the earlier hours of the night (Figs. IA and 3). the neutral sugar fractions from two early and two late light period pineapple leaf samples. Assays indicated between 15 Diurnal Changes in Glucans ptg and 40 ,ug of glucose and sucrose were applied in the various aliquots. The chromatogram was developed by de- The pattern of diurnal changes in total glucans was the scending solvent flow at 22°C in 4:1:5 butanol:acetic inverse of changes in malic acid (Fig. 1, A and C). Starch acid:water. Sugars were detected with aniline oxalate, 1% comprised 70% to 85% ofthe total glucan fraction throughout 94 CARNAL AND BLACK Plant Physiol. Vol. 90,1989

from total glucans was relatively constant (-2.9 ,gmol hexose eq/gfw. h) during the first 5 h ofthe dark period, but decreased 140 -A dramatically to -0.4 ,.mol hexose eq/gfw. h later in the dark period (Figs. 1C and 3). Early in the dark period hexose 0110 turnover from low mol wt glucans accounted for approxi- Ei100.- mately two-thirds of the decline in total glucan (Fig. 3). After TITRATABLE A the first 2 h ofdarkness, however, the turnover ofhexose from ACID starch accounted for all hexose turnover from total glucans. During the last 3 h of the dark period decreases in the starch 60 pool were countered by increases in lower mol wt glucan LU pools (Fig. 3) indicating starch conversion to smaller polymers.

Fructosans No fructose polymers were detected. LU-

120 Diurnal Changes in Neutral Ethanol-Soluble Sugars ~,MALATE The three sugars detected in neutral ethanol-soluble sugar fractions of pineapple leaves co-chromatographed with fruc- ~40 tose, glucose, and sucrose (4). The dynamics of diurnal changes in pool sizes of these sugars were subsequently ex- amined. Differences in absolute total soluble sugar levels (herein defined as the sum of fructose, glucose, and sucrose levels) and in pool sizes of component soluble sugars among 680 ©- plants were evident (Fig. 2); these between-plant differences were maintained throughout the 24 h period. Hence, turn- overs of total soluble sugars, fructose, and glucose during the dark period were very consistent, as evidenced by the low SE

0~~~~~~~~~ for the mean turnover of these components (Table I). Some 60 ~ ~ ~ between-plant variability was attributed to different prehisto- ries of the two sets of plants examined. The three plants exhibiting higher hexose levels, but lower sucrose levels (Fig. 40 ~~~~~~~ 2, B-D) were moved from the greenhouse to the environment LU chamber in the fall; the three plants with lower hexose but higher sucrose concentrations were placed in the chamber and acclimated to standard conditions in the spring of the following year. Glucose and fructose concentrations changed reciprocally with changes in malic acid levels (Figs. IA and 2, B and C). Sucrose concentration, however, remained relatively constant 20 STARCH the when FiurLU 1.Po0ief()tirtbeai,()mai Ncd C oa during light period malic acid levels were declining A~~~a rapidly (Figs. IA and 2D). A small but consistent decline in 1 4 7 10 1 47 10 sucrose was noted early in the dark period. Thereafter, sucrose levels increased to predark levels at the same time the tissue TIME OF DAY was accumulating malic acid (Figs. IA and 2D). Fructose content ofleaftissues ranged from 2.5 to 3.5 times Figure 1. Pool sizes of (A) titratable acid, (B) malic acid, (C) total larger than glucose content. Nearly equivalent amounts of glucan, and (D) starch in pineapple leaves during a 24-h period. Data hexose, however, were depleted from each of these pools are given for four to six plants. Titratable acids are given as Amol the B and C; mean rate malate eq/gfw (conversion based on 2 geq acid per Amol malate). during night (Fig. 2, Table I). The of Starch and total glucans (starch plus low mol wt glucans) are given decline of total soluble sugars was lower early and late in the as ,mol hexose eq/gfw. Dark bars indicate the 9-h dark period. dark period (-4.7 and 3.8 gmol hexose eq/gfw h, respec- tively), and highest from 2 to 5 h into the night (5.5 ,umol hexose eq/gfw .h) (Figs. 2A, 3). the 24 h period and changes in total glucan levels were primarily attributable to changes in starch (Fig. 1, C and D). Stoichiometry of Carbohydrate and Malic Acid Changes Starch levels accounted for 83% ofthe variance in total glucan levels while changes in lower mol wt glucan pools explained A 1:1 stoichiometry of hexose depletion from soluble-sugar only 6% of the variance (n = 42). The rate of hexose turnover pools and hexose required to support malic acid accumulation CARBOHYDRATE RESERVES DURING CAM 95

Table I. Increase in Malic Acid and Decrease in Various Carbohydrate Fractions in Pineapple Leaves during a 9-h Dark Period Fraction Change, Mean ± SE (n) 'rnol/gfw Malate 84.0 ± 3.3 (5) Total glucan 18.6 ± 6.4 (4) Starch 15.6 ± 2.5 (4) Total soluble sugars 41.8 ± 1.2 (5) Glucose 19.1 ± 1.2 (5) Fructose 20.2 ± 1.3 (5) Sucrose 2.5 ± 1.0 (5)

at night (1 hexose:2 malic acid) was observed (Table I). In four of five plants examined hexose turnover from soluble ~TOTAL sugar pools was completely sufficient to account for malic SOLUABLE acid accumulation. In the fifth plant, soluble sugar turnover 14 * could supply 91% of the triose-P required for conversion to PEP. In contrast to the very consistent turnover in total soluble sugars at night, considerable variability in turnover of glucan pools was noted among plants (Table I, Fig. 3). On the average, hexose derived from glucans could account for ap- proximately 44% ofthe malic acid that accumulated. Overall, s 1Q0fX hexose depletion from all carbohydrate pools examined ex- ceeded hexose required to support malic acid synthesis by ELLI 1401; _Ib A a . about 40%.

100A Rates of Glycolysis and Activities of PPi-PFK and ATP- PFK LU RUCTOSE The maximum rate of malic acid accumulation occurred co IA 60 during the last 3-h sampling period in the dark (Figs. and 3). The mean maximum rate of malic acid accumulation of ~~1 0 7 1 about 14 ,tmol/gfw * h (Table II) required glycolytic catabolism of hexose to proceed minimally at a rate of approximately 7 4 0A ,umol hexose/gfw. h. The measured mean maximum rate of hexose depletion from all carbohydrate pools of 9.4 ,imol GLUCOSE hexose/gfw- h occurred between 2 and 5 h into the dark period (Table II). These two estimates of the mean maximal rates of glycolysis over a 3-h period during the acidification phase of Liur 2. 40Polszso-()ttlslbluas(umo rcoe CAM were considered to represent the range of mean maxi- A~~~~~ mal rates of glycolysis occurring at night. The mean substrate-saturated activities (±SE) of PPi-PFK and ATP-PFK from pineapple leaves at 30°C were 290 ± 13.9 itmol Fru-1,6-P2 produced/gfw.h (n = 5) and 16 ± 1.3 umol ~SUCROSE Fru-1,6-P2/gfw.h (n = 5), respectively. Qio values (±SE) for PPi-PFK and ATP-PFK determined from sequential assays at 30°C and about 15°C were 1.86 ± 0.06 (n = 5) and 1.79 ± 10 1 4 7 10 1 4 7 10 0.04 (n = 4), respectively. TIME OF DAY

Figure 2. Pool sizes of (A) total soluble sugars (sum of fructose, DISCUSSION glucose, and sucrose); (B) fructose; (C) glucose; and (D) sucrose in pineapple leaves during a 24-h period. Data are given for six plants. Malic Acid For three plants (A, , V) fructose was determined for neutral (Dowex In pineapple, as in other CAM plants, malic acid is the treated) extracts (see "Materials and Methods"). Data for all soluble acid that shows dramatic in concentration sugar fractions are given as Mmol hexose eq/gfw. Dark bars indicate organic changes the 9-h dark period. during CAM (Fig. 1, A and B) (20). Citrate levels in pineapple leaves may change diurnally; however, the direction ofchange varies with respect to changes in malic acid (21, 23). Isocitrate 96 CARNAL AND BLACK Plant Physiol. Vol. 90,1989

c3)

80 0) o O10 0 0 60 E o sE

- w 40

C) r 20

o

-1 to 2 2to5 5to8 -1 to 8 DARK PERIOD TIME INTERVAL (hr)

Figure 3. Increase in pool size of malic acid (Amol/gfw) and decreases in pool sizes of various carbohydrates (given as ,mol triose eq/gfw) in pineapple leaves during the dark phase of CAM. Data (mean ± SE) are pool size changes occurring during three 3-h time intervals beginning 1 h before the onset of the dark period (-1 h) and ending 8 h into the dark period. Cumulative pool size changes from -1 h to 8 h are also presented. For low mol wt glucans a value less than "zero" denotes an increase in the pool size during the time interval indicated. EM, malic acid (n = 5); OS, total soluble sugars (sum of fructose, glucose, and sucrose) (n = 5); C1G, total glucan (starch plus low mol wt glucans) (n = 4). MSt, starch (n = 4); MGs, low mol wt glucans (n = 4).

is detectable only at low concentrations and changes in keto Table II. Mean Maximum Ratesa of Malic Acid Accumulation and acids contribute little to changes in titratable acidity (21). Carbohydrate Depletion in Pineapple Leaves during the Dark Phase of CAM Glucans as Carbohydrate Sources for Malate Synthesis Fraction Mean Maximum Rate of Change, Mean + SE (n) It is generally accepted in the literature on CAM that glucan ,imol malic acid or hexose/gfw/h pools constitute the sources of hexose that support CO2 ac- Malic acid 14.0 ± 0.8 (5) cumulation at night. This is clearly not the case for pineapple. Total soluble sugars 6.1 ± 0.6 (5) Although glucan pools in pineapple leaves fluctuate recipro- Total glucan 3.8 ± 1.2 (4) cally with malic acid levels (Fig. 1, C and D), the turnover of Total carbohydrate 9.4 ± 2.0 (4) hexose from these pools could, on the average, supply only a Mean maximum rate is mean hourly rate of change during the 3 about 44% of the hexose required to support malic acid h period in the dark during which maximum turnover of the selected synthesis (Table I). Aloe arborescens, a CAM plant that, like fraction occurred. pineapple, uses PEP carboxykinase as the major enzyme for malate decarboxylation, likewise shows insufficient turnover CARBOHYDRATE RESERVES DURING CAM 97 of hexose from glucan pools to account for malic acid accu- to CAM (18). The relative contributions of various soluble mulation (28). In contrast, hexose depletion from either starch sugar and glucan pools to malic acid formation at night alone or starch plus low mol wt glucans is sufficient to account remains to be determined, but it is clear that soluble sugar for all the malic acid that accumulates in species of Crassu- pools provide the majority, if not all, triose-P required. laceae (22, 25, 32) and in Opuntia aurantiaca (22). Hexose depletion from starch reserves in pineapple leaves Phosphofructokinase Activities and Estimated Rates of accounts for over 80% ofthe decrease in total glucans at night Glycolysis (Table I). Net removal of hexose from low mol wt glucan The turnover of fructose and glucose pools in pineapple pools occurs only in the early hours of the dark period (Fig. leaves to provide hexose for conversion to PEP, and hence 3). These pools increase later in the dark period at the expense malate, is significant with respect to the compartmentation of of larger mol wt polymers, i.e. starch. Starch pools previously glycolytic carbon flow during acidification and the involve- reported for pineapple leaves showed little diurnal variation ment of PPi-PFK and ATP-PFK in hexose catabolism during (23). Our calcuations using data of Sideris et al. (23) indicate CAM. Starch degradation to triose-P occurs within the chlo- that only about 2% ofthe increase in titratable acidity reported roplast (assuming permeability features of chloroplast enve- could be accounted for by removal and processing of hexose lopes are similar in CAM and C3 species). Fru-6-P generated from starch reserves. The differences in these results may be in the chloroplast is therefore inaccessible to PPi-PFK and its due to differences in techniques used to extract and solubilize phosphorylation to Fru- 1,6-P2 must occur via the chloroplast starch. Starch pools in plants used in the present study were isozyme of ATP-PFK. The mobilization of monosaccharides considerably larger than those determined by Sideris et al. from vacuoles at night, however, provides a source of hexose (23). Alternatively, differences in growth conditions in the to the cytoplasm, thereby increasing the likelihood that PPi- two studies may have influenced the relative allocations of PFK, a cytoplasmically localized enzyme (2, 4, 7, 19), is carbon to starch and sucrose. Indeed, higher sucrose levels involved in hexose catabolism coupled to malic acid accu- and a larger turnover of sucrose in pineapple leaves at night mulation. It is perhaps noteworthy that species of Crassula- were obserbed by Sideris et al. (23) than in our study. Since ceae that couple starch degradation to malic acid accumula- very little sucrose is located in protoplasts of pineapple me- tion exhibit PPi-PFK activities that are either similar to or sophyll cells (18), the larger sucrose turnover observed by lower than ATP-PFK activities, while PPi-PFK activities in Sideris et al. (23) indicates that larger amounts of carbohy- pineapple are about 15 to 20 times higher than activities of drate were being exported. ATP-PFK (5). Substrate-saturated activities of pineapple leaf PPi-PFK at Neutral Soluble Sugar Pools as Carbohydrate Sources 30°C are 30 to 40 times the estimated mean maximum rates for Malate Synthesis ofglycolysis at night in pineapple (Table II). Activities oftotal Pineapple leaves contain large soluble sugar pools (150- leaf ATP-PFK at 30°C range from just sufficient to about 250 ,umol hexose eq/gfw) that fluctuate diurnally in parallel twice the activity required to account for rates of malic acid with glucan fluctuations and reciprocally with changes in accumulation. Several factors, however, make it unlikely that malic acid levels (Figs. 1 and 2A). The 1:2 stoichiometry of sufficient ATP-PFK activity is present in pineapple leaves to hexose turnover from soluble sugar pools and malic acid catalyze the phosphorylation of Fru-6-P at rates required accumulation (or 1:1 stoichiometry of triose-P derived from during the acidification phase of CAM. First, acidification is soluble hexose reserves and malic acid) at night suggests that promoted by low night temperatures. Plants in this study hexose turnover from these reserves is strongly coupled to experienced a night temperature of 15°C. The Qio values malate biosynthesis (Table I, Fig. 3). The data of Sideris et al. determined for PPi-PFK and ATP-PFK indicate that activities (23), when extrapolated to the end ofthe dark period, likewise of these enzymes at 15°C are 39% and 42% (i.e. 114 and 6.7 indicate that the depletion of soluble sugar reserves is suffi- pmol Fru-1,6-P2/min .gfw) of the respective catalytic capaci- cient to account for malic acid accumulation. In contrast to ties observed at 30°C. Therefore, at 15°C ATP-PFK activity these data for pineapple, soluble sugar pools in Bryophyllum is nearly sufficient to account for rates of hexose turnover tubiflorum and Kalanchoe daigremontiana are small (-5 required for malic acid accumulation, but alone, is insufficient ,umol/gfw) and remain constant throughout the day/night to account for rates oftotal hexose turnover. PPi-PFK activity cycle (25). Furthermore, although hexose derived from "1C- at 15°C, however, is an order of magnitude higher than that labeled starch supports malate biosynthesis, it is essentially required to account for observed rates of hexose turnover. excluded from flow through soluble sugar pools (25). Soluble Second, neither PPi-PFK nor the isozymes of ATP-PFK sugar pools in B. calycinum, while larger than in the latter are expected to be substrate saturated in vivo. Metabolite two species (30 ,umol/gfw), behave inconsistently with respect levels in a few CAM plants have been reported (e.g. Ref. 6), to changes in malic acid (22, 29), again indicating that these however, levels of metabolites in various subcellular com- pools are not coupled to malic acid accumulation. partments of CAM plants are not known. Fru-2,6-P2 concen- Given the magnitude of soluble sugar concentrations in trations in pineapple leaves increase at night (9) to levels that pineapple leaves it is unlikely that these sugars are sequestered are consistent with full activation of pineapple PPi-PFK (4). in the cytosol or smaller organelles. Vacuoles isolated from Reported Fru-6-P concentrations in the cytoplasm and chlo- pineapple leaves contain substantial amounts of glucose and roplasts of C3 plants (e.g. Ref. 31) are near the apparent Km fructose-and diurnal changes in vacuolar hexose levels are (Fru-6-P) for Fru-2,6-P2-activated pineapple PPi-PFK (0.35 consistent with the interpretation that these pools are coupled mM) (4) and the apparent Km (Fru-6-P) for ATP-PFK from 98 CARNAL AND BLACK Plant Physiol. Vol. 90, 1989

Table Ill. Energy Requirements for Malic Acid Synthesis and Accumulation during the Dark Phase of CAM: Starch as the Hexose Donor, ATP-PFK as the Catalyst for the First Committed Reaction of Glycolysisa Starch (Glc,) + P04-2 Glc-1 -P-2 + Glc,-l Glc-1 -P-2 Glc-6-p-2 Glc-6-P-2 4 1- Fru-6-p-2 Fru-6-p-2 + ATP-4 4 Fru-1,6-P2 4 + ADP-3 + H+ Fru-1 ,6-P2-4 4 G3p-2 + DHAp-2 DHAP-2 G3p-2 2 G3p-2 + 2 Pi2 + 2 NAD4 2 1,3-bisPGA-4 + 2 NADH + 2 H4 2 1,3-bisPGA-' + 2 ADP-3 4 0 2 3-PGA-3 + 2 ATP-4 2 3-PGA-3 4 2 2-PGA-3 2 2-PGA-3 4 2 PEP-3 + 2 H20 2 C02 + 2H20 2 HC03- + 2 H4 2 PEP-3 + 2 HC03- 4 2 OAA-2 + 2 P04-2 2 OAA-2 + 2 NADH + 2 H4 0 2 Malate 2 + 2 NAD4 2 ATP-4 2 ADP-3 + 2 P04-2 + 2 H4 2 Malate2cY, + 4 H+,y, 2 Malic acidvac Net: 1 Glc + 2 C02 + ATP-4 2 Malic acidvac + ADP-3 + pi-2 + H4 Net energy per mol malic acid accumulated: 0.5 ATP 0.5 ADP + 0.5 Pi + 0.5 H+ a DHAP, dihydroxyacetone phosphate; PGA, phosphoglycenc acid; OAA, oxaloacetic acid; cyt, cytoplasm; vac, vacuole.

Table IV. Energy Requirements for Malic Acid Synthesis and Accumulation during the Dark Phase of CAM: Free Hexose as the Hexose Donor, PPi-PFK as the Catalyst for the First Committed Reaction of Glycolysisa Hexose + ATP-' Hexose-W2 + ADP-3 + H+ Fru-6-p-2 + PPi-' Fru-1.6-P2 ' + pi-2 Fru-1,6P2' > G3 + UHAP'! DHAp-2 > G3p-2 2 G3p-2 + 2 pi-2 + 2 NAD+ 2 1 ,3-bisPGA-4 + 2 NADH + 2 H4 2 1,3-bisPGA-4 + 2 ADP-3 2 3-PGA-3 + 2 ATP-4 2 3-PGA-3 2 2-PGA-3 2 2-PGA-3 > 2 PEP-3 + 2 H20 2 C02 + 2 H20 > 2 HCO3 + 2 H+ 2 PEP3 + 2 HC03- 2 OAA2 + 2 P042 2 OAA-2 + 2 NADH + 2 H* 2 Malate-2 + 2 NAD+ 2 ATP-4 2 ADP-3 + 2 P04-2 + 2 H4 2 Malate 24, + 4 H-acidcy> 2 Malic Net: hexose + 2 C02 + ATP4 + PPi-4 2 Malic acidvac + ADP3 + 3 pj2 + H4 Net energy per mol malic acid accumulated: 0.5 ATP + 0.5 PPi > 0.5 ADP + 1.5 Pi + 0.5 H4 a Abbreviations as in Table Ill.

CAM and C3 plants (16, 17, 26). PPi concentrations reported sufficient to account for the rate of malic acid accumulation. for pineapple leaves (14 nmol/gfw) (24) are near the apparent PPi-PFK activity, however, is sufficient to support glycolysis Km (PPi) of 16 AM for Fru-2,6-P2-activated pineapple PPi- at rates required. PFK (4). ATP concentrations in chloroplasts and in the A third consideration regarding the ability of ATP-PFK to cytoplasm ( 14, 31) are 5 to 10 times higher than apparent Km sustain glycolysis at rates observed in pineapple at night is the (ATP) values reported for ATP-PFK for C3 or CAM plants relative proportion of hexose derived from soluble sugar ver- (17, 26). Based on these estimates of substrate concentrations, sus glucan pools. If extrachloroplastic free hexose is the sole ATP-PFK and PPi-PFK would at best catalyze the phospho- source ofhexose for malic acid accumulation, then ATP-PFK rylation of Fru-6-P at 25% to 50% of the maximum velocity is even less capable of catalysis at required rates because only predicted. Therefore, at 15°C and with reasonable estimates about 45% of the ATP-PFK activity is localized in the cyto- of substrate concentrations, ATP-PFK activity is clearly in- plasm of pineapple mesophyll cells (4). CARBOHYDRATE RESERVES DURING CAM 99

Models for Glycolysis Supported by Starch Reserves that required for acidification by about 40% (Table I). A small versus Free Hexose Reserves during CAM proportion of this excess will sustain respiration at rates reported for CAM plants. The remainder of the carbohydrate Luttge et al. (20) have argued against hexose pools serving turnover is channeled to biosynthesis. PPi is perhaps gener- as hexose sources for PEP formation during the acidification ated in sufficient quantity as a product of metabolite activa- phase ofCAM in two species ofKalanchoe. These researchers tion reactions associated with the synthesis of sucrose and conclude that predicted rates of ATP production in these various polymers. species can not meet energy requirements for malate biosyn- thesis and transport into the vacuole unless polysaccharides are the source of hexose for acidification and hexose moieties CONCLUSIONS are released by phosphorolysis. Table III outlines the energet- Neutral soluble sugar pools serve as carbohydrate reserves ics of malic acid production and accumulation in the vacuole providing hexose to support malic acid accumulation during when starch is degraded by phosphorolysis and ATP-PFK CAM in pineapple. In pineapple leaves, levels of glucose and catalyzes the phosphorylation of Fru-6-P. The net energy fructose fluctuate in parallel with starch and reciprocally with requirement to convert 0.5 mol hexose-P to 1 mol malate and respect to malic acid (Figs. 1 and 2). Soluble sugar turnover accumulate 1 mol malate and 2 mol H+ in the vacuole is 0.5 at night is very consistent among plants and is sufficient to mol ATP. One-half mol ADP, 0.5 mol Pi and 0.5 mol H+ are account for all hexose required to support malic acid accu- the extravacuolar products (Table III). Input of hexose into mulation (Table I). Total glucan turnover at night is less glycolysis via hexokinase instead of starch phosphorylase consistent among plants and could account, on the average, would increase the ATP requirement to 1 mol ATP per 1 mol for about 44% of the malic acid that accumulates. malic acid accumulated. Since the average rate of malic acid A strong case can be made for the involvement of PPi-PFK accumulation in K. tubiflora is nearly twice the predicted rate in hexose turnover during the acidification phase of CAM in of ATP production (20), the model depicted in Table III is pineapple. First, a source of hexose-P is available to this the only model involving ATP-PFK consistent with respira- cytoplasmic enzyme. Second, the catalytic capacity of PPi- tory rates and rates of malic acid accumulation. PFK at low temperatures that promote acidification and The model presented in Table IV outlines the energetics of under non-substrate saturating conditions is many fold in malic acid accumulation when free hexose pools donate hex- excess of rates of glycolysis required to support malic acid ose for triose-P production and PPi-PFK catalyzes the phos- accumulation, energy and/or other biosynthetic needs. The phorylation of Fru-6-P. One-half mol ATP and 0.5 mol PPi catalytic capacity of ATP-PFK in pineapple is not sufficient are consumed per mol malic acid accumulated in the vacuole. to meet these needs. Third, the energetics of malic acid One and one-half mol Pi, 0.5 mol ADP, and 0.5 mol H+ are accumulation at the expense of free hexoses are consistent generated in the cytoplasm. If respiratory rates in pineapple with dark respiration rates reported for CAM plants (1, 20) if are similar to those in K. tubiflora, night-time acidification PPi-PFK but not ATP-PFK catalyzes the phosphorylation of rates in pineapple can be supported by turnover of hexose Fru-6-P. from glucose and fructose pools as long as PPi-PFK, but not ATP-PFK, catalyzes the phosphorylation of Fru-6-P and LITERATURE CITED either a sufficiently large PPi pool exists or PPi production is 1. Andre M, Thomas DA, von Willert DJ, Gerbaud A (1979) not at the expense of ATP generated oxidatively at night. If Oxygen and carbon dioxide exchanges in Crassulacean acid neither of the latter conditions is met, the ratio of rates of metabolism plants. Planta 147: 141-144 dark respiration to malic acid accumulation must be higher 2. Black CC, Carnal NW, Kenyon WH (1982) Compartmentation and the regulation of CAM. In IP Ting, M Gibbs, eds, Cras- in pineapple than in Kalanchoe. sulacean Acid Metabolism. American Society of Plant Phys- The model in Table IV predicts that when energy require- iologists, Rockville, MD, pp 51-68 ments for malic acid accumulation balance ATP production 3. Block RJ, Durrum EL, Zweig G (1958) A Manual of Paper 1 mol Pi is generated per mol malic acid accumulated in the Chromatography and Paper Electrophoresis. Academic Press, New York vacuole. The model in Table III predicts no net Pi change 4. Carnal NW (1984) Pyrophosphate: 6-phosphofructokinase in under the same circumstance. Interestingly, Daley and Vines plants: discovery, characterization, and an examination of its (8) report that Pi levels in pineapple leaves increase at night, role in carbohydrate metabolism. PhD dissertation, University although not with the stoichiometry predicted if all hexose of Georgia, Athens, GA 5. Carnal NW, Black CC (1983) Phosphofructokinase activities in utilized for malic acid production were derived from soluble photosynthetic organisms: The occurrence of pyrophosphate hexose reserves. Overall, the malic acid to Pi stoichiometry dependent 6-phosphofructokinase in plants and algae. Plant would be modulated by the degree to which starch is utilized Physiol 71: 150-155 for malate production and/or ATP-PFK is involved in gly- 6. Cockburn W, McAulay A (1977) Changes in metabolite levels in Kalanchoe daigremontiana and the regulation of malic acid colytic processing of hexose for malate synthesis. accumulation in Crassulacean acid metabolism. Plant Physiol Although we cannot estimate the percentage of carbohy- 59: 455-458 drate turnover occurring via PPi-PFK, the data strongly sup- 7. Cseke C, Weeden NF, Buchanan BB, Uyeda K (1982) A special port the involvement ofthis enzyme in carbohydrate process- fructose bisphosphate functions as a cytoplasmic regulatory metabolite in green leaves. Proc Natl Acad Sci USA 79: 4322- ing during acidification. Further work is therefore required to 4326 address the question of the source of PPi for this enzyme. 8. Daley LS, Vines HM (1977) Diurnal fluctuations of inorganic Carbohydrate turnover at night in pineapple leaves exceeds orthophosphate in pineapple (Ananas comosus (L.) Merr.) 100 CARNAL AND BLACK Plant Physiol. Vol. 90,1989

leaves and a possible role of ATP-ase. Plant Science Lett 10: of malate accumulation in the vacuoles of Kalanchoe tubiflora 289-298 cells. FEBS Lett 126: 81-84 9. Fahrendorf T, Holtum JAM, Mukherjee U, Latzko E (1987) 21. Milburn TR, Pearson DJ, Ndegwe NA (1968) Crassulacean acid Fructose 2,6-bisphosphate, carbohydrate partitioning, and metabolism under natural tropical conditions. New Phytol 67: Crassulacean acid metabolism. Plant Physiol 84: 182-187 883-897 10. Garland WJ, Dennis DT (1980) Plastid and cytosolic phospho- 22. Pucher GW, Vickery HB, Abrahams MD, Leavenworth CS fructokinases from developing endosperm of Ricinus commu- (1949) Studies on the metabolism of Crassulacean plants: nis. Arch Biochem Biophys 204: 302-309 diurnal variation of organic acids and starch in excised leaves 11. Graham D, Smydzuk J (1965) Use ofanthrone in the quantitative ofBryophyllum calycinum. Plant Physiol 25: 610-620 determination of hexose phosphates. Anal Biochem 11: 246- 23. Sideris CP, Young HY, Chun HHQ (1948) Diurnal changes and 255 growth rates as associated with ascorbic acid, titratable acidity, 12. Gutmann I, Wahlefeld AW (1974) L-(-)-Malate: determination carbohydrate and nitrogenous fractions in the leaves ofAnanas with malate dehydrogenase and NAD. In HU Bergmeyer, ed, comosus (L.) Merr. Plant Physiol 23: 38-69 Methods of Enzymatic Analysis, Vol 3. Academic Press, New 24. Smyth DA, Black CC (1984) Measurement of pyrophosphate York, pp 1585-1589 content of plant tissues. Plant Physiol 75: 862-864 13. Hassid WZ, Neufeld EF (1964) Quantitative determination of 25. Sutton BG (1975) The path of carbon in CAM plants at night. starch in plant tissues. In RL Whistler, RJ Smith, JN BeMiller, Aust J Plant Physiol 2: 377-387 eds, Methods in Carbohydrate Chemistry, Vol 4. Academic 26. Sutton BG (1975) Glycolysis in CAM plants. Aust J Plant Physiol Press, New York, pp 33-36 2: 389-402 14. Heber U (1974) Metabolite exchange between chloroplasts and 27. Turner JP, Turner DH (1980) The regulation of glycolysis and the pentose phosphate pathway. In DD Davies, ed, The Bio- cytoplasm. Annu Rev Plant Physiol 25: 393-421 chemistry of Plants, Vol 2, Chap 7. Academic Press, New 15. Huber SC (1982) Biochemical basis for partitioning ofphotosyn- York, pp 279-316 thetically fixed carbon between starch and sucrose in soybean 28. Verbucheln 0, Steup M (1984) Carbon metabolism and malate (Glycine max Merr.) leaves. Plant Physiol 69: 1404-1406 formation in the CAM plant Aloe arborescens. In C Sybesma, 16. Kelly GJ, Latzko E (1977) Chloroplast phosphofructokinase. I. ed, Advances in Photosynthesis Research, Vol 3. Martinus Proof of phosphofructokinase activity in chloroplasts. Plant Nijhoff, The Hague, pp 421-424 Physiol 60: 290-294 29. Vickery HB (1954) The effect of abnormally prolonged alterna- 17. Kelly GJ, Latzko E (1977) Chloroplast phosphofructokinase. II. ting periods of light and darkness upon the composition of Partial purification, kinetic and regulatory properties. Plant Bryophyllum calycinum leaves. Plant Physiol 29: 520-526 Physiol 60: 295-299 30. Whiting BH, van de Venter HA, Small JGC (1979) Crassulacean 18. Kenyon WH, Severson RF, Black CC (1985) Maintenance carbon acid metabolism in jointed cactus (Opuntia aurantiaca) Lin- cycle in Crassulacean acid metabolism plant leaves. Plant dley. Agroplantae 11: 41-43 Physiol 77: 183-189 31. Wirtz W, Stitt M, Heldt HW (1980) Enzymic determination of 19. Kruger NJ, Kombrink E, Beevers H (1983) Pyrophosphate Fruc- metabolites in the subcellular compartments of spinach pro- tose 6-phosphate phosphotransferase in germinating castor toplasts. Plant Physiol 66: 187-193 bean seedlings. FEBS Lett 153: 409-412 32. Wood WML (1952) Organic acid metabolism of Sedum praeal- 20. Luttge U, Smith JAC, Marigo G, Osmond CB (1981) Energetics tum. J Exp Bot 3: 336-355